Everyone wants to feel special, but it turns out you really are: Your brain reacts and responds to stimuli so uniquely to you that scientists can use a “brain print” to identify you from others with 100 percent accuracy. This new study of brain biometrics out of Binghamton University (BU) was published in the journal IEEE Transactions on Information Forensics and Security.
Researchers hooked 56 people up to an electroencephalogram headset while the participants viewed a series of 400 images: 100 sine gratings (wavy lines), 100 low frequency words, 100 images of foods, and 100 celebrity faces. Each image flashed onscreen for only 200 milliseconds, a process that took about 30 minutes. Most of the images were black and white, though some were in color—and participants had to respond to the color images by hitting a button. “The goal was to try and find imagery that we thought had a chance of being responded to really uniquely from person to person,” Sarah Laszlo, assistant professer of psychology at BU, tells mental_floss. “Almost by intuition we thought a couple of those kinds of things would be pictures of celebrities and food. Then we had another category of visual stimuli that look just like ripples because those strongly activate the visual cortex in humans.”
It turns out that the visual cortex of each human brain is in many ways unique, like a fingerprint. “Each person has a slightly different set of folds in their visual cortex, so we knew if we were punching the visual cortex, it would be very different from person to person,” Laszlo says.
This study built upon one in 2015, published in Neurocomputing, that was able to identify one person out of a group of 32 by that person’s responses to a set of words, with a rate of 97 percent accuracy.
“When I first got into this I thought we’d be lucky if we could do [identify a person] 25 percent of the time,” Laszlo says. When their 100 percent accuracy rate turned up, she says, “My jaw dropped.”
She attributes the success rate to including data inspired by a psychological approach: “This was the first attempt that had psychologists on the team instead of only engineers, and I think that turned out to be really beneficial.” Still, she gives credit to lead engineer Zhanpeng Jin, assistant professor of electrical and computer engineering, for “being willing to get help and think that someone knew something he didn’t know.”
The actual identification process requires a computer’s assistance to read the brain prints, and it has strong implications for building a new form of data security.
She references a case of Chinese cyber-espionage in which an agent stole a database full of fingerprints of 2.5 million federal workers. “So now any system that was secured by those fingerprints was compromised permanently because fingerprints don’t change and you can’t grow new fingers,” she says. “These need to be replaced with something more secure. We think that a brain print might be able to do that.”
Their work is promising because it could be extremely hard to hack. And even if it could be hacked, one could easily create a new brain print. “If your brain print was a response to food and celebrities, like we’ve done here, you could cancel it and replace it with a new one that is a response to plants and vegetables, or something different,” Laszlo says.
They are now working on ways to hack the process to see if it is possible, for example, to impersonate someone’s brain print. “To do that, we are doing this awesome experiment where people come in, we call them brain hackers, and we non-invasively stimulate their brain in the pattern of someone else’s brain, to see whether we can make their brain activity more similar to somebody else’s,” she says.